Tools for evaluating risk of terrorist acts using fuzzy sets and belief/plausibility

Terrorist acts are intentional and therefore differ significantly from "dumb" random acts that are the subject of most risk analyses. There is significant epistemic (state of knowledge) uncertainty associated with such intentional acts, especially for the likelihood of specific attack scenarios. Also, many of the variables of concern are not numeric and should be treated as purely linguistic (words). Epistemic uncertainty can be addressed using the belief/ plausibility measure of uncertainty, an extension of the traditional probability measure of uncertainty. Fuzzy sets can be used to segregate a variable into purely linguistic values. Linguistic variables can be combined using an approximate reasoning rule base to map combinations of fuzzy sets of the constituent variables to fuzzy sets of the resultant variable. We have implemented the mathematics of fuzzy sets, approximate reasoning, and belief/plausibility into Java software tools. The PoolEvidence
^copy
software tool combines evidence (pools) from different experts. The LinguisticBelief
^copy
software tool evaluates the risk associated with scenarios of concern using the pooled evidence as input. The tools are not limited to the evaluation of terrorist risk; they are useful for evaluating any decision involving significant epistemic uncertainty and linguistic variables. Sandia National Laboratories' analysts have applied the tools to: risk of terrorist acts, security of nuclear materials, cyber security, prediction of movements of plumes of hazardous materials, and issues with nuclear weapons. This paper focuses on evaluating the risk of acts of terrorism.